QANTAS Cyber Incident

[odysseus]

How gullible do they think we are to say they formatted it but didn't draft it with AI?

If they can draft the letter then the punctuation is not an issue. AI had more involvement in that letter than just formatting.

 

[Seat0B]

Just this April I received $570 from a class action against ANZ Credit Cards... it must have taken many years as I have no recollection (nothing in email either) of ever joining a class action... once I actually got the money I instantly regretted not joining the Optus one... if the lawyers took 40% before my cut then wow they must have made out like bandits. The lesson? Don't ever change your email address

Not even if it’s been hacked!

 

[kookaburra75]

How odd. I checked a ChatGPT conversation I had which included a bulleted list and it produces standard bullet points, not generated images of bullets. Even copying and pasting into Word brings it across as a standard list, not with custom bullet points.

That's my experience too. I found the article a bit odd. I suspect it might be how the formatting as handled by their bulk email system, rather than using MS Word

 

[Daver6]

I don't see any issue with AI being used for this. Done properly, it's not just asking AI to write an apology. There would be some proper prompting and several iterations plus potentially some manual editing done too.

 

[SeaWolf]

I don't see any issue with AI being used for this. Done properly, it's not just asking AI to write an apology.

And the AI will definitely be more compassionate and empathetic than QF staff.

 

[TheInsider]

It's like these that are reporting the AI thing don't use AI

 

[SeatBackForward]

I suspect it might be how the formatting as handled by their bulk email system, rather than using MS Word

Apparently it was meta data associated with the dot point image that identified it as being generated by ChatGPT - Why someone was looking at this is equally mind boggling, but yes it may well be that those style of dot points somehow became part of a template used in their bulk email system.

But having said that, they've admitted using ChatGPT also.

 

[Seat0B]

My last class action netted me nearly 5 figures from VW Dieselgate. .

Yes we also did well on that one, but not that well!

 

[kookaburra75]

Apparently it was meta data associated with the dot point image that identified it as being generated by ChatGPT - Why someone was looking at this is equally mind boggling, but yes it may well be that those style of dot points somehow became part of a template used in their bulk email system.
But having said that, they've admitted using ChatGPT also.

I saw that report, which didn't really tell us much - but I digress (but isn't that what AFF is about? )

It's like these that are reporting the AI thing don't use AI

So true. But for those who get really excited about that sort of thing, there is this advice

I found 5 AI content detectors that can correctly identify AI text 100% of the time

I've been testing AI content detectors for two years now. They're getting more and more reliable. These are the ones you can count on.

www.zdnet.com

 

[RooFlyer]

I don't see any issue with AI being used for this. Done properly, it's not just asking AI to write an apology. There would be some proper prompting and several iterations plus potentially some manual editing done too.

Fair enough, but if you are writing a sincere apology, surely you have someone sit down and write what they, or the company feel. Have some professional writer polish it by all means - but asking a machine to do it for you?

To me, its sort of like saying, in your car "Hey Siri (or Google etc). Send a text of condolence to my friend X including my sympathies for their recent loss". You could do it, but would you?

It's like these that are reporting the AI thing don't use AI

They aren't apologising to millions for their cough-up.

 

[sudoer]

They aren't apologising to millions for their cough-up.

Yes, context matters tremendously here...

Edit: I'll add that I understand both viewpoints. I respect the viewpoint of those who don't care about the use of AI - fair enough. But I also absolutely understand the view of those who put off by AI having any hand in the drafting of a 'sincere apology', and think it can detract from the perceived sincerity of the message. Even if only 10% of recipients feel that way, that's 570,000 reasons not to use AI.

 

[DejaBrew]

Yes, context matters tremendously here...

Edit: I'll add that I understand both viewpoints. I respect the viewpoint of those who don't care about the use of AI - fair enough. But I also absolutely understand the view of those who put off by AI having any hand in the drafting of a 'sincere apology', and think it can detract from the perceived sincerity of the message. Even if only 10% of recipients feel that way, that's 570,000 reasons not to use AI.

I don't disagree with you, but at the same time, I think that Qantas was on a hiding to nothing regardless of their approach. Give AI some hand in creating the comms, and it's insincere. VH pens the entire thing herself, and people would find a million and one other reasons to criticise. You only need to look at the response from customers to the emails when they believed it was penned by VH/Qantas staff without any ChatGPT involvement.

I'm not saying that I would have used AI if I were in Qantas' position, but I can certainly understand why they might have leveraged it to help refine the finished product. And realistically, if it wasn't for the fact that someone clearly had way too much time on their hands, most of us would have been none the wiser.